Changeset 17045 for branches/2952_RF-ModelStorage/HeuristicLab.Algorithms.DataAnalysis

Timestamp:

06/27/19 15:46:20 (5 years ago)

Author:

mkommend

Message:

#2952: Intermediate commit of refactoring RF models that is not yet finished.

Location:

branches/2952_RF-ModelStorage/HeuristicLab.Algorithms.DataAnalysis/3.4

Files:

• GradientBoostedTrees/GradientBoostedTreesAlgorithm.cs (modified) (1 diff)
• GradientBoostedTrees/GradientBoostedTreesModel.cs (modified) (2 diffs)
• GradientBoostedTrees/GradientBoostedTreesModelSurrogate.cs (modified) (2 diffs)
• HeuristicLab.Algorithms.DataAnalysis-3.4.csproj (modified) (1 diff)
• RandomForest/RandomForestClassification.cs (modified) (3 diffs)
• RandomForest/RandomForestModel.cs (modified) (8 diffs)
• RandomForest/RandomForestModelFull.cs (added)
• RandomForest/RandomForestModelSurrogate.cs (added)
• RandomForest/RandomForestRegression.cs (modified) (4 diffs)
• RandomForest/RandomForestUtil.cs (modified) (2 diffs)

branches/2952_RF-ModelStorage/HeuristicLab.Algorithms.DataAnalysis/3.4/GradientBoostedTrees/GradientBoostedTreesAlgorithm.cs

@@ -269 +269 @@
 
         if (ModelCreation == ModelCreation.SurrogateModel) {
-          model = new GradientBoostedTreesModelSurrogate(problemData, (uint)Seed, lossFunction, Iterations, MaxSize, R, M, Nu, (GradientBoostedTreesModel)model);
+          model = new GradientBoostedTreesModelSurrogate((GradientBoostedTreesModel)model, problemData, (uint)Seed, lossFunction, Iterations, MaxSize, R, M, Nu);
         }
 

branches/2952_RF-ModelStorage/HeuristicLab.Algorithms.DataAnalysis/3.4/GradientBoostedTrees/GradientBoostedTreesModel.cs

@@ -34 +34 @@
   // this is essentially a collection of weighted regression models
   public sealed class GradientBoostedTreesModel : RegressionModel, IGradientBoostedTreesModel {
-    // BackwardsCompatibility3.4 for allowing deserialization & serialization of old models
-    #region Backwards compatible code, remove with 3.5
-
     [Storable(Name = "models")]
     private IList<IRegressionModel> __persistedModels {
@@ -53 +50 @@
       get { return weights; }
     }
-    #endregion
 
     public override IEnumerable<string> VariablesUsedForPrediction {

branches/2952_RF-ModelStorage/HeuristicLab.Algorithms.DataAnalysis/3.4/GradientBoostedTrees/GradientBoostedTreesModelSurrogate.cs

@@ -90 +90 @@
     private Func<IGradientBoostedTreesModel> CreateLazyInitFunc(IGradientBoostedTreesModel clonedModel) {
       return () => {
-        return clonedModel == null ? RecalculateModel() : clonedModel;
+        return clonedModel ?? RecalculateModel();
       };
     }
 
     // create only the surrogate model without an actual model
-    public GradientBoostedTreesModelSurrogate(IRegressionProblemData trainingProblemData, uint seed,
+    private GradientBoostedTreesModelSurrogate(IRegressionProblemData trainingProblemData, uint seed,
       ILossFunction lossFunction, int iterations, int maxSize, double r, double m, double nu)
       : base(trainingProblemData.TargetVariable, "Gradient boosted tree model", string.Empty) {
@@ -106 +106 @@
       this.m = m;
       this.nu = nu;
+
+      actualModel = new Lazy<IGradientBoostedTreesModel>(() => RecalculateModel());
     }
 
     // wrap an actual model in a surrograte
-    public GradientBoostedTreesModelSurrogate(IRegressionProblemData trainingProblemData, uint seed,
-      ILossFunction lossFunction, int iterations, int maxSize, double r, double m, double nu,
-      IGradientBoostedTreesModel model)
+    public GradientBoostedTreesModelSurrogate(IGradientBoostedTreesModel model, IRegressionProblemData trainingProblemData, uint seed,
+      ILossFunction lossFunction, int iterations, int maxSize, double r, double m, double nu)
       : this(trainingProblemData, seed, lossFunction, iterations, maxSize, r, m, nu) {
       actualModel = new Lazy<IGradientBoostedTreesModel>(() => model);

branches/2952_RF-ModelStorage/HeuristicLab.Algorithms.DataAnalysis/3.4/HeuristicLab.Algorithms.DataAnalysis-3.4.csproj

@@ -394 +394 @@
     <Compile Include="RandomForest\RandomForestClassification.cs" />
     <Compile Include="RandomForest\RandomForestModel.cs" />
+    <Compile Include="RandomForest\RandomForestModelFull.cs" />
     <Compile Include="RandomForest\RandomForestRegression.cs" />
     <Compile Include="RandomForest\RandomForestRegressionSolution.cs" />
+    <Compile Include="RandomForest\RandomForestModelSurrogate.cs" />
     <Compile Include="RandomForest\RandomForestUtil.cs" />
     <Compile Include="SupportVectorMachine\SupportVectorClassification.cs" />

branches/2952_RF-ModelStorage/HeuristicLab.Algorithms.DataAnalysis/3.4/RandomForest/RandomForestClassification.cs

@@ -21 +21 @@
 
 using System.Threading;
+using HEAL.Attic;
 using HeuristicLab.Common;
 using HeuristicLab.Core;
@@ -26 +27 @@
 using HeuristicLab.Optimization;
 using HeuristicLab.Parameters;
-using HEAL.Attic;
 using HeuristicLab.Problems.DataAnalysis;
 
@@ -144 +144 @@
 
       if (CreateSolution) {
-        var solution = new RandomForestClassificationSolution(model, (IClassificationProblemData)Problem.ProblemData.Clone());
+        var solution = model.CreateClassificationSolution(Problem.ProblemData);
         Results.Add(new Result(RandomForestClassificationModelResultName, "The random forest classification solution.", solution));
       }

branches/2952_RF-ModelStorage/HeuristicLab.Algorithms.DataAnalysis/3.4/RandomForest/RandomForestModel.cs

@@ -23 +23 @@
 using System.Collections.Generic;
 using System.Linq;
+using HEAL.Attic;
 using HeuristicLab.Common;
 using HeuristicLab.Core;
 using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding;
-using HEAL.Attic;
 using HeuristicLab.Problems.DataAnalysis;
 using HeuristicLab.Problems.DataAnalysis.Symbolic;
@@ -34 +34 @@
   /// Represents a random forest model for regression and classification
   /// </summary>
-  [StorableType("A4F688CD-1F42-4103-8449-7DE52AEF6C69")]
+  [Obsolete("This class only exists for backwards compatibility reasons. Use RFModelSurrogate or RFModelFull instead.")]
+  [StorableType("9AA4CCC2-CD75-4471-8DF6-949E5B783642")]
   [Item("RandomForestModel", "Represents a random forest for regression and classification.")]
   public sealed class RandomForestModel : ClassificationModel, IRandomForestModel {
@@ -139 +140 @@
     public IEnumerable<double> GetEstimatedValues(IDataset dataset, IEnumerable<int> rows) {
       double[,] inputData = dataset.ToArray(AllowedInputVariables, rows);
-      AssertInputMatrix(inputData);
+      RandomForestUtil.AssertInputMatrix(inputData);
 
       int n = inputData.GetLength(0);
@@ -157 +158 @@
     public IEnumerable<double> GetEstimatedVariances(IDataset dataset, IEnumerable<int> rows) {
       double[,] inputData = dataset.ToArray(AllowedInputVariables, rows);
-      AssertInputMatrix(inputData);
+      RandomForestUtil.AssertInputMatrix(inputData);
 
       int n = inputData.GetLength(0);
@@ -175 +176 @@
     public override IEnumerable<double> GetEstimatedClassValues(IDataset dataset, IEnumerable<int> rows) {
       double[,] inputData = dataset.ToArray(AllowedInputVariables, rows);
-      AssertInputMatrix(inputData);
+      RandomForestUtil.AssertInputMatrix(inputData);
 
       int n = inputData.GetLength(0);
@@ -315 +316 @@
 
       alglib.dfreport rep;
-      var dForest = CreateRandomForestModel(seed, inputMatrix, nTrees, r, m, 1, out rep);
+      var dForest = RandomForestUtil.CreateRandomForestModel(seed, inputMatrix, nTrees, r, m, 1, out rep);
 
       rmsError = rep.rmserror;
@@ -353 +354 @@
 
       alglib.dfreport rep;
-      var dForest = CreateRandomForestModel(seed, inputMatrix, nTrees, r, m, nClasses, out rep);
+      var dForest = RandomForestUtil.CreateRandomForestModel(seed, inputMatrix, nTrees, r, m, nClasses, out rep);
 
       rmsError = rep.rmserror;
@@ -361 +362 @@
 
       return new RandomForestModel(problemData.TargetVariable, dForest, seed, problemData, nTrees, r, m, classValues);
-    }
-
-    private static alglib.decisionforest CreateRandomForestModel(int seed, double[,] inputMatrix, int nTrees, double r, double m, int nClasses, out alglib.dfreport rep) {
-      AssertParameters(r, m);
-      AssertInputMatrix(inputMatrix);
-
-      int info = 0;
-      alglib.math.rndobject = new System.Random(seed);
-      var dForest = new alglib.decisionforest();
-      rep = new alglib.dfreport();
-      int nRows = inputMatrix.GetLength(0);
-      int nColumns = inputMatrix.GetLength(1);
-      int sampleSize = Math.Max((int)Math.Round(r * nRows), 1);
-      int nFeatures = Math.Max((int)Math.Round(m * (nColumns - 1)), 1);
-
-      alglib.dforest.dfbuildinternal(inputMatrix, nRows, nColumns - 1, nClasses, nTrees, sampleSize, nFeatures, alglib.dforest.dfusestrongsplits + alglib.dforest.dfuseevs, ref info, dForest.innerobj, rep.innerobj);
-      if (info != 1) throw new ArgumentException("Error in calculation of random forest model");
-      return dForest;
-    }
-
-    private static void AssertParameters(double r, double m) {
-      if (r <= 0 || r > 1) throw new ArgumentException("The R parameter for random forest modeling must be between 0 and 1.");
-      if (m <= 0 || m > 1) throw new ArgumentException("The M parameter for random forest modeling must be between 0 and 1.");
-    }
-
-    private static void AssertInputMatrix(double[,] inputMatrix) {
-      if (inputMatrix.ContainsNanOrInfinity())
-        throw new NotSupportedException("Random forest modeling does not support NaN or infinity values in the input dataset.");
     }
 

branches/2952_RF-ModelStorage/HeuristicLab.Algorithms.DataAnalysis/3.4/RandomForest/RandomForestRegression.cs

@@ -20 +20 @@
 #endregion
 
+using System.Collections.Generic;
+using System.Linq;
 using System.Threading;
+using HEAL.Attic;
 using HeuristicLab.Common;
 using HeuristicLab.Core;
@@ -26 +29 @@
 using HeuristicLab.Optimization;
 using HeuristicLab.Parameters;
-using HEAL.Attic;
 using HeuristicLab.Problems.DataAnalysis;
 
@@ -144 +146 @@
 
       if (CreateSolution) {
-        var solution = new RandomForestRegressionSolution(model, (IRegressionProblemData)Problem.ProblemData.Clone());
+        var solution = model.CreateRegressionSolution(Problem.ProblemData);
         Results.Add(new Result(RandomForestRegressionModelResultName, "The random forest regression solution.", solution));
       }
     }
+
 
     // keep for compatibility with old API
@@ -157 +160 @@
     }
 
-    public static RandomForestModel CreateRandomForestRegressionModel(IRegressionProblemData problemData, int nTrees,
-      double r, double m, int seed,
-      out double rmsError, out double avgRelError, out double outOfBagRmsError, out double outOfBagAvgRelError) {
-      return RandomForestModel.CreateRegressionModel(problemData, nTrees, r, m, seed,
-        rmsError: out rmsError, avgRelError: out avgRelError, outOfBagRmsError: out outOfBagRmsError, outOfBagAvgRelError: out outOfBagAvgRelError);
+    public static RandomForestModelFull CreateRandomForestRegressionModel(IRegressionProblemData problemData, int nTrees,
+     double r, double m, int seed,
+     out double rmsError, out double avgRelError, out double outOfBagRmsError, out double outOfBagAvgRelError) {
+      return CreateRandomForestRegressionModel(problemData, problemData.TrainingIndices, nTrees, r, m, seed, out rmsError, out avgRelError, out outOfBagRmsError, out outOfBagAvgRelError);
+    }
+
+    public static RandomForestModelFull CreateRandomForestRegressionModel(IRegressionProblemData problemData, IEnumerable<int> trainingIndices, int nTrees, double r, double m, int seed,
+    out double rmsError, out double avgRelError, out double outOfBagRmsError, out double outOfBagAvgRelError) {
+
+      var variables = problemData.AllowedInputVariables.Concat(new string[] { problemData.TargetVariable });
+      double[,] inputMatrix = problemData.Dataset.ToArray(variables, trainingIndices);
+
+      alglib.dfreport rep;
+      var dForest = RandomForestUtil.CreateRandomForestModel(seed, inputMatrix, nTrees, r, m, 1, out rep);
+
+      rmsError = rep.rmserror;
+      outOfBagRmsError = rep.oobrmserror;
+      avgRelError = rep.avgrelerror;
+      outOfBagAvgRelError = rep.oobavgrelerror;
+
+      return new RandomForestModelFull(dForest, problemData.TargetVariable, problemData.AllowedInputVariables);
+
+      //return RandomForestModel.CreateRegressionModel(problemData, nTrees, r, m, seed,
+      //rmsError: out rmsError, avgRelError: out avgRelError, outOfBagRmsError: out outOfBagRmsError, outOfBagAvgRelError: out outOfBagAvgRelError);
     }
 

branches/2952_RF-ModelStorage/HeuristicLab.Algorithms.DataAnalysis/3.4/RandomForest/RandomForestUtil.cs

@@ -27 +27 @@
 using System.Linq.Expressions;
 using System.Threading.Tasks;
+using HEAL.Attic;
 using HeuristicLab.Common;
 using HeuristicLab.Core;
 using HeuristicLab.Data;
 using HeuristicLab.Parameters;
-using HEAL.Attic;
 using HeuristicLab.Problems.DataAnalysis;
 using HeuristicLab.Random;
@@ -89 +89 @@
 
   public static class RandomForestUtil {
+    public static void AssertParameters(double r, double m) {
+      if (r <= 0 || r > 1) throw new ArgumentException("The R parameter for random forest modeling must be between 0 and 1.");
+      if (m <= 0 || m > 1) throw new ArgumentException("The M parameter for random forest modeling must be between 0 and 1.");
+    }
+
+    public static void AssertInputMatrix(double[,] inputMatrix) {
+      if (inputMatrix.ContainsNanOrInfinity())
+        throw new NotSupportedException("Random forest modeling does not support NaN or infinity values in the input dataset.");
+    }
+
+    internal static alglib.decisionforest CreateRandomForestModel(int seed, double[,] inputMatrix, int nTrees, double r, double m, int nClasses, out alglib.dfreport rep) {
+      RandomForestUtil.AssertParameters(r, m);
+      RandomForestUtil.AssertInputMatrix(inputMatrix);
+
+      int info = 0;
+      alglib.math.rndobject = new System.Random(seed);
+      var dForest = new alglib.decisionforest();
+      rep = new alglib.dfreport();
+      int nRows = inputMatrix.GetLength(0);
+      int nColumns = inputMatrix.GetLength(1);
+      int sampleSize = Math.Max((int)Math.Round(r * nRows), 1);
+      int nFeatures = Math.Max((int)Math.Round(m * (nColumns - 1)), 1);
+
+      alglib.dforest.dfbuildinternal(inputMatrix, nRows, nColumns - 1, nClasses, nTrees, sampleSize, nFeatures, alglib.dforest.dfusestrongsplits + alglib.dforest.dfuseevs, ref info, dForest.innerobj, rep.innerobj);
+      if (info != 1) throw new ArgumentException("Error in calculation of random forest model");
+      return dForest;
+    }
+
+
     private static void CrossValidate(IRegressionProblemData problemData, Tuple<IEnumerable<int>, IEnumerable<int>>[] partitions, int nTrees, double r, double m, int seed, out double avgTestMse) {
       avgTestMse = 0;